combust. Sci. and Tach., 1994, Vol. 103, pp.

41-61 0 1994 OPA (Overseas Publishers Association)

Reprints available directly from the publisher Amsterdam B.V. Published under license by
photocopying permitted by license only Gordon and Breach Science Publishers SA
Printed in Malaysia

Sources of Combustion lrreversibility

W. R. DUNBAR and N . LlOR Department of Mechanical Engineering

and Applied Mechanics, University of Pennsylvania,
Philadelphia, PA 19104-6315

(Received June 2,1992; in final form October 5, 1994)

ABSTRACT- Approximately 113of the useful energy of the fuel is destroyed during the combustion process
used in electrical power generation. This study is an attempt to clarify and categorize the reasons for the
exergy destruction taking place in combustion processes. The entropy production is separated into three
subprocesses: (1) combined diffusion/fuel oxidation, (2) "internal thermal energy exchange" (heat transfer),
and (3) the product constituent mixing process. Four plausible process paths are proposed and analyzed. The
analyses are performed for two fuels: hydrogen and methane. The results disclose that the majority (about
314) of the exergy destruction occurs during the internal thermal energy exchange. The fuel oxidation, by
itself, is relatively efficient, having an exergetic efficiency of typically 94% to 97%.

Key Words: Combustion, exergy, second-law analysis, power generation, irreversibility,thermodynamics

NOTATION

a,, Specific chemical exergy, kJ/kgrnole

Specific flow exergy, kJ/kgmole
a,, Specific thermal mechanical exergy, kJ/kgrnole
Molar specific heat, kJ/kgmole K
Convective energy rate, kJ/s

Volumetric rate of entropy production, kJ/K.m3s

Specific enthalpy, kJ/kgmole
Molar flux of component i, kgmole/m2s
Molar flow rate, kgmole/s
Pressure, kPa
Atmospheric pressure, kPa
Heat transfer rate, kJ/s
Universal gas constant, kJhgmo1e.K
Rate of reaction j, kgmolels
Specific entropy, k ~ h g m o l e . ~
Entropy production rate, kJ1K.s
Temperature, K
Reference temperature, K

41
 42 W. R. DUNBAR AND N. LIOR

A Exergy rate, kJ/s

& Exergy destruction rate, kJ/s
A Chemical affinity, kJ/kgmole
Pi Chemical potential of component i, (kJ/kgmole fuel)
Q Spatial thermal conductivity vector, kJ1m.K.s
T Stress tensor, N/m2
Xi Mole fraction of component i

Subscripts
i Mass stream index
j, n Species index
o Dead state conditions

Superscripts
Rate (per unit time)

INTRODUCTION

Past studies have revealed the combustion process, of the many processes occurring in
the typical electricity-producing power plant, as the single largest contributor of exergy
losses (cf. Gaggioli et al., 1975). With present technology, fuel oxidation by conven-
tional ("uncontrolled") combustion at atmospheric pressure consumes about 113 of the
fuel's utilizable energy. The objective of this work is to investigate the sources of this
irreversibility and its underlying reasons.
Specifically, the approach taken in this study is to (i) describe and quantify the overall
(global) entropy production taking place during combustion, and (ii) separate and
quantify the amount of entropy production associated with the subprocesses of
combustion, namely constituent mixing, oxidation, and internal thermal energy ex-
change, along four conceivable representative paths of the global combustion process.
This computation along prescribed process paths, proposed by Dunbar and Lior
(1990), is an alternative to the extremely difficult rigorouos solution of the full field and
state equations (Navier-Stokes, energy, entropy generation, and thermodynamic prop-
erties, cf. Gaggioli 1961, 1962) combined in a combustion process with mass transfer
and reaction kinetics equations, all tightly coupled. Rigorous analysis of combustion
can be performed (cf. Buckmaster and Ludford, 1982; Arai et al., 1986) but due to the
many simplifications that are introduced to ease the mathematical problem, and the
many uncertainties, the accuracy of the result is not likely to exceed that of the
simplified solution shown here. The "rigorous" analysis has indeed so far only been
applied to the simplest heatlmass transfer problems, such as flow in two-dimensional
channels without any chemical reactions (Bejan, 1979; San et al., 1987;Poulikakos and
Johnson, 1989), and the case of premixed flames stabilized above a flat-flame burner
(Arpaci and Selamet, 1988),and even that required the use of a number of simplifica-
tions and empirical correlations.
* *..
 COMBUSTION IRREVERSIBILITY 43

To conclude, while it is widely known by now that combustion creates a major

exergy loss, it was not yet made clear how and where specifically this loss is incurred in
this highly-complex process. Such quantitative understanding is sought in this study.
Apart from its fundamental value, this understanding is a vital starting point for the
search for practical means for the realization of more efficientcombustion and power
generation systems.

GLOBAL ANALYSIS

General description
We consider steady combustion in a well-insulated combustion chamber. The term
"global" refers here to consideration of the combustor as a single "black-box" control
volume, with conditions known or determined only at the control-volume inlet and
outlet. The global modeling is performed here first, and the equilibrium state of the
combustor products, the extent of reaction (defined as the molar amount of fuel reacted
per mole of fuel input), and the effects of excess air are determined, from balances
on energy and chemical species, property relations, and the relevant temperature-
dependent equilibrium constants. Two analyses are performed: one with hydrogen and
one with methane as the fuel. The fuels are assumed to be pure, at the ambient
conditions of 25"C, 1 atm. These conditions also define the dead (reference)state; for the
hydrogen and methane fuels it is that oftheir combustion products: H,O, and CO, and
H,O, respectively. The oxidant is atmospheric air with an assumed composition of
21% oxygen and 79% nitrogen. The amount of excess air is varied, ranging from
0%-100%.
In the case of hydrogen combustion, the product gas stream is assumed to consist of
unreacted diatomic hydrogen and oxygen, inert diatomic nitrogen, and water. This
assumption is made even though there may in reality be additional species present at
the calculated adiabatic flame temperatures, due to molecular dissociation, incomplete
combustion, and other reactions. For example, at a temperature of 2500 K, there is
a very small amount (less than 0.1 %) of molecular dissociation of diatomic hydrogen,
oxygen and nitrogen to the respective monatomic species, as well as similarly small
amounts of dissociation of H,O (cf. Vargaftik, 1975; Wark, 1977).The amount of OH-
and N O molecules formed is somewhat higher. While the inclusion of these additional
reactions would make the results of the analysis slightly more precise, they would not
have a significant effect on the objectives of this study, which is the understanding of the
major sources of irreversibility in combustion, and would add much to the complexity
of the analysis. Such minor side-reactions were therefore ignored in this study.
Similarly, in the case of methane combustion, the reactions are described by

CH, + ZO, -+ CO + 2H20,

 44 W. R. DUNBAR AND N. LIOR

with inert nitrogen, carried in with the combustion air and removed with the combus-
tion ~roducts.existing in both.

Second Law analysis

I Due to the adiabatic boundary restriction and the fact that no work is produced during
conventional steadv adiabatic combustion. the exerav balance is

in out

i.e., the exergy associated with the entering matter is equal to the exergy of the exiting
matter plus the irreversible destruction of exergy associated with the combustion. Here,
the total flow exergy of stream i is

where the total specific exergy of a given flow stream (e.g., stream i), as summed over all
the species j, is

'fi = CXijafij,
i

where

The specific flow exergies are evaluated by employing calculational procedures found
in the literature (cf. Rodriguez, 1980), by first separating them into their thermom-
echanical components

where the specific flow exergy, af, is composed of two exergy contributions: (1) the
specific thermomechanical exergy, a,, and (2) the specific chemical exergy, a,,.
Assuming ideal gas behavior for all components,

I Here for all chemical speciesj in gas stream i the enthalpies are thus expressed as
PT
1I
1

I and the entropies are expressed as

pi
sj = sjo + c p j F - Rln -.
Po
 COMBUSTION IRREVERSIBILITY 45

The exergy destruction rate can also be expressed as

and it was computed in this study in two ways: by using Eq. (4) and then by using
Eq. (12), to double-check the correctness of the results.
The entropy for Eq. (12) was calculated from the balance

Enthalpies of formation, absolute entropies, chemical exergies, and ideal gas heat
capacity coefficients were obtained from Reynolds and Perkins (1977), Rodriguez
(1980), Sonntag and Van Wylen (1982), and Gurvich and Veyts (1989).

Boundary conditions
The boundary conditions fur the global, steady reactor are: (1)the fuel and air entrance
temperatures are at the assumed reference, ambient temperature of 25"C, (2) incoming
fuel and air compositions, (3) the product gas stream exits the reactor under chemical
equilibrium conditions, (4) all gas streams are at atmospheric pressure, and (5) the
combustion chamber (reactor) walls are adiabatic.

Results ofthe global analysis

Results of the global analysis of hydrogen combustion are contained in Figure 1. Based
on the equilibrium reaction equations, the extent of reaction basically becomes 1.00
(implyingcomplete oxidation of fuel) above 50% excess air. The equilibrium, adiabatic,

0 20 40 60 80 100
Excess Air (%)
FIGURE 1 Exergetic Efficiency, Extent of Reaction, Product Temperature and Irreversibility, vs. E x q s
Air for Global Analysis of Hydrogen Combustion.
 46 W. R. DUNBAR AND N. LIOR

product gas temperature ranges from 1646 K to 2433 K, decreasing with an increase in
the amount of excess air. The amount of entropy production during the combustion
increases with excess air, ranging from 53,667 to 78,833 kJ/(kgmole H,). Finally, the
concomitant thermodynamic cost of this entropy production is determined by evaluat-
ing the exergy destruction (combustion irreversibility) displayed in Figure 1 via the
exergy efficiency, as a function of the amount of excess air. The exergy efficiency is
defined as the ratio of exergy outputs to exergy inputs. The exergetic efficiency of
hydrogen combustion ranges from 66% to 77%, decreasing with increasing amount of
excess air. The irreversibility is calculated by Equation (12). In conclusion, such
conventional combustion destroys approximately 23% to 34% of the useful energy of
hydrogen fuel for the investigated range of excess air.
Results of the global analysis of methane combustion (Eqs. 1 and 2) are contained in
Figure 2. For the range of excess air amount studied, the extent of reaction for water
formation ranges from 0.981 to 1.00, basically becoming 1.00 (implying complete
formation of water) above 40% excess air. The extent of reaction for carbon dioxide
formation ranges from 0.909 to 1.00, basically becoming 1.00 above 50% excess air.
The equilibrium, adiabatic, product gas temperature ranges from 1480K to 2249 K,
decreasing with an increase in the amount of excess air. The amount of entropy
production increases with increasing amount of excess air, ranging from (2.456)105 to
(3.611)105kJ/(kgmole CH,). The exergetic efficiency of methane combustion ranges
from 60% to 72%, decreasing with increasing amount of excess air. In other words,
conventional combustion destroys approximately 28% to 40% of the useful energy of
methane fuel for the investigated range of excess air.
The significant degradation of the potential to produce useful work during combus-
tion, computed above to be 23% to 40% with hydrogen and methane as fuels, was also
observed for other types of hydrocarbon fuel combustion (cf. Hedman et al., 1980).This

0 20 40 60 80 100

Molar Amount after

Contplete Combustion

C02 Amount per

. Molar Amount after

60 1 ' 1 . 1 - 1 . 1200
0 20 40 60 80 160
Excess Air (%)
FIGURE 2 Exergetic Efficiency, Extent of Reaction, Product Temperature and Irreversibility vs. Excess
Air for Global Analysis of Methane Combustion.
*..
 cOMBUSTION IRREVERSIBILITY 47

irreversibility has significant detrimental impact in all combustion-based energy

conversion systems in the energy industry, such as residential and commercial space
heating furnaces, industrial furnaces for materials processing, combustion chambers
within power plant boilers, external combustion engines, and internal combustion
engines (in which the irreversibilities are even higher). It is thus important to improve
our understanding of the phenomena contributing to entropy generation in combus-
tion, and we approach it in the following by a breakdown and quantification of the
subprocesses taking place during combustion.
The initial condition before fuel combustion (state 1)'includesa fuel stream and an air
stream, both at ambient temperature and pressure. The final condition, that after fuel
combustion (state 2), is characterized by (i) the presence of a number of new chemical
species (e.g., water, carbon dioxide), (ii) a higher temperature, and (iii) all constituents
being mixed. The processes which cause these physical changes: reactant mixing, fuel
oxidation, heat transfer, and product mixing, produce the aforementioned amounts of
irreversible entropy production. The immediate objective is to conceptually separate
and quantify the amounts of entropy production associated with these various
physicochemical subprocesses of combustion.

BREAKDOWN O F ENTROPY PRODUCTION IN COMBUSTION

Hirschfelder, Curtiss and Bird (1954) present an expression for the local rate of entropy
production (discussion of the breakdown of the exergy equation into its components
can be found in Dunbar et al., 1992). Integrating the equation over the volume of the
system results in a relation for the global rate of entropy production:

{-(0.v T) - (TVV) - 1[ji.V(pi - p,)] + i l l j ~+j 1(ji.Fi))dV

i i
(14)

The first term on the right-hand-side is the contribution to entropy production due to
heat transfer, the second term is that due to momentum transfer (fluid friction),the third
and fifth terms are those due to the action of chemical potential gradients and body
forces, respectively, on diffusion mass fluxes, and the fourth term is due to chemical
reactions. The volume integrals in Equation (18) are process-dependent and hence,
path-dependent in the thermodynamic sense. To evaluate these volume integrals, the
approach taken here is to assume hypothetical sequences of subprocesses, wherein
these volume integrals may then be evaluated by surface integrals.
In this study it is assumed that fluid friction and body-force energy are negligible and
that thus four physicochemical subprocesses are the primary contributors to the global
rate of entropy production: (i) reactant diffusion, (ii) reaction (fuel oxidation), (iii)
internal thermal energy exchange (i.e., heat transfer between gas constituents within the
reactor), and (iv) product mixing. Because the magnitudes of the various contributions
to entropy production are path-dependent, four conceivable process-paths are studied
and the results thereof are compared. Two of these paths assume incremental stages of
 48 W. R. DUNBAR AND N. LIOR

fuel oxidation; two assume instantaneous fuel oxidation. ?'he equations used are those
described in the Global Analysis section above.
The combustion process is envisioned here to proceed in the following manner.
the air and fuel enter the reactor, the oxygen and fuel molecules, which have an intrinsic
affinity for each other, are "drawn" together in a diffusion process which consumes
useful power, used to separate them from the reactant gas stream.
Having approached the fuel, and with the possibility of a number of series/parallel
steps, the oxygen reacts with the fuel, forming product molecules. During this process,
there is a concomitant energy conversion: net changes of energy in the forms of (1)
"internal-chemical energy" - energy associated with intramolecular forces, (2) radi-
ation energy, and (3) "internal-thermomechanical energy" - associated with particle
motions and intermolecular forces between system constituents (Hirschfelder et al.,
1954).Having stabilized (reacted),these interactions between the participating species
have thus added to the system entropy, destroying potential useful power.
At the instant following the exothermic reaction, the product molecules leave the
reaction site with a kinetic and photon energy which is much higher than the statistical
average energy of the neighboring particles which have not reacted yet. Momentum
(and associated kinetic energy)and radiative transfers therefore take place, bringing the
reaction product temperature down to an equilibrium value consistent with that of the
surrounding medium. This process, which is here called an "internal thermal energy
exchange", also consumes useful power, increasing the entropy of the system. Finally, in
a process which occurs simultaneously with the internal-thermal energy exchange, but,
consistent with the approach in this study, separated here for the sake of insight, the
product molecules mix with the remaining global system constituents, further reducing
the system exergy. Hence, the above description provides a categorization (breakdown)
of subprocesses that occur within the global combustion process. To summarize, the
processes are theoretically separated as follows:
(1) a diffusion process where the oxygen and fuel molecules are drawn together),
(2) a chemical reaction (oxidation of the fuel),
(3) an internal-thermal energy exchange, where the product molecule "shares" its
kinetic energy with its neighbors, and
(4) a mixing process whereby system constituents mix uniformly.
The contributions to exergy destruction are calculated below for four process paths.
The paths were chosen to represent somewhat limiting conditions in the determination
of subprocess irreversibilities, and to be physico-chemically plausible.

Path 1
In this process path the breakdown of combustion irreversibility is investigated by
assuming that (i) fuel oxidation occurs in stages of reaction (i.e., not instantaneously),
and (ii) fuel combustion takes place in the subprocess order af (a) reactant diffusion
combined with fuel oxidation, (b) internal thermal energy exchange, and (c) product
mixing.
Figure 3 displays schematically the technique of analysis. For example, in chamber 1,
*.
the oxygen (only the incremental stoichiometric amount which will react in this
 COMBUSTION IRREVERSIBILITY

I
I First Incremental I Second Incremental I
! Extent of Reaction ! Extent of Reaction !
Unreacted I /

Fuel Unreactcd Fuel / 'j!

Feed Fuel
* Q
* Q i 'jl
I:: 1
A
- -'
H2
I il;,l

1 i(
A -q ... P 111:
- -'-q O2
$. -
Q
Global Reador
Combustion i',!!
1,:

Air
A
I + (Stoich.
Amount)
Produas Under
Chemical
Feed ,Amount)
O2
(Stoich.
Q Equilibrium
Conditions
Chamber 1 Chamber 3 Chamber 5
Chamber 2 Chamber 4 Chamber 6

Chamber
1
2
process
-
'Diffusion/Reaction (H2+ 1R02 H *0)
Internal Thermal Energy Exchange
3 Mix (ProductsDepleted Air)
4 Diffusion/Reaction
5 Internal Thermal Energy Exchange
6 Mix (ProductsDepleted Air)

FIGURE 3 Hypothetical Combustion Chamber for Process Path 1.

chamber) enters a compartment where it reacts with the incremental amount of fuel
consumed. Upon reaction, the products exit into chamber 2 at the temperature
characterized by a completed reaction in an adiabatic chamber. Flowing in a separate
compartment, isolated from the reaction momentarily, is the amount of unreacted fuel
and the oxygen-depleted air/product streams.
In chamber 2, heat transfer (but no mass transfer) is allowed to take place between the
gas particles in the separated compartments of this chamber. All gas temperatures
exiting chamber 2 are the same (i.e., thermal equilibrium is assumed). Thus, the
additional boundary conditions for chamber 2 are: (1) heat transfer (but no mixing)
allowed between the compartments within the reactor (but no heat transfer to ambient
surroundings), and (2) thermal equilibrium between all constituents at the exit of
chamber 2.
Finally, the products from the reaction in chamber 1 and the unreacted air
components mix in chamber 3. The boundary conditions for chamber 3 are: (I) uniform
mixture of all reaction products and oxygen-depleted air stream at exit, and (2)
adiabatic boundaries.
Following the first incremental extent of reaction, the gas components flow into
chamber 4, repeating these steps of combustion for the second (and subsequent)
incremental extent (s) okeaction. This procedure is repeated, with consequent gradual
increase in temperature, until the gas constituents reach the fuel ignition temperature
[assumed here to be 582C for hydrogen and 690C for methane, typical values for
atmospheric pressure combustion of these fuels, taken from Babcock and Wilcox,
19781, whereupon the fuel oxidation is assumed to be instantaneous (i.e., all the
remaining fuel is oxidized immediately at the fuel ignition temperature). This repetition
till ignition, and not till the equilibrium temperature, which is higher, is more realistic,
 50 W. R. DUNBAR A N D N . LIOR

since heat transfer following oxidation tends to diminish the temperature as compared
with that predicted by equilibrium. The final products of combu~tionthen exit the
reactor under the required global system equilibrium conditions.

EIffects of increment size

An analysis was first performed to determine the sensitivity of the results to the number
of increments chosen for the process described in Figure 3. These results are displayed
in Figures 4 and 5 for the cases of 0% and 50% excess air amount.
At a number of increments of about 30 the results approach a value which does not
change with the increase of the number of increments. This convergence occurs because
the changes in the conditions of the reactants diminish in each cell-increment as the
number of increments increases, for given overall process beginning and end states. For
example, when 500 increments are chosen, the increase in temperature in each
increment is only about 1 to 2 K (in the computed case of 100% excess air).
The exergy destruction due to oxidation decreases with increasing number of
increments because larger percentages of the fuel oxidation are evaluated under more
efficient conditions; i-e., at higher temperatures and therefore closer to equilibrium
(hence more reversible see Fig. 6). Five hundred increments were chosen for the
following combustion irreversibility breakdown analysis, thus assuring the determina-
tion of increment-size-independent exergy destruction values for the assumed sequence
of subprocesses in this path.

Breakdown of overall exergy consumption

The relevant results of the breakdown of exergy losses in hydrogen combustion are
given in Figure 7. The significant result is that the largest subprocess exergetic

-
1 Overall (~~obal)----\, L
50000 - - 50000
-
z Internal Thermal
40000- Energy Exchange - 40000
2c
.s 30000 - - 30000

Number of Increments
FIGURE 4 Exergy Destruction vs. Number of Process Increments for 0% Excess Air. Fuel: H,.
** .
 COMBUSTION IRREVERSIBILITY

0 100 200 300 400 500

42. Internal Thermal

Energy Exchange
+
.d
40000 - 40000
0
2
*-'
V)

B
a
ZS Fuel Oxidation
7
I" '

0 100 200 300 400 500

Number of Increments
FIGURE 5 Exergy Destruction vs. Number of Process Increments for 50% Excess Air. Fuel: H,.

88 88
0 400 800 1200 1600 2000
Reactants Temperature (K)

I FIGURE 6 Exergetic Efficiency of Hydrogen Oxidation vs. Reactants Temperature.

consumption takes place during the internal thermal energy exchange (chambers 2,5,
etc. in Fig. 3). The exeregetic efficiency of this subprocess is 73% to 83%, decreasing
with increasing amounts of excess air, i.e. approximately 72-77% of the overall exergy
loss of the combustion process is associated with it. The reaction (oxidation, in
chambers 1,4, etc., in Fig. 3) has a 94% to 95% exergetic efficiency and constitutes
about 15% to 18% of the total exergy loss, and gas constituent mixing (in chambers
 W. R. DUNBAR AND N. LIOR

Internal Thermal

-
\\ Oxidation
7Mixing

0 I ' I I I 0
0 20 40 60 80 100
Excess Air ( O h )
FIGURE 7 Hydrogen Combustion Subprocess Exergy Destruction vs. Excess Air, Path 1.

3,6, etc., in Fig. 5) has an exergetic efficiency of 96.5% to 97.4%, constituting the
reamining 8 % to 10% of the total exergy loss.
The overall (global)exergetic efficiency ranges from 66.5% to 77.3%, decreasing with
increasing amounts of excess air.

Path 2
In this process path, described in Figure 8, we study the breakdown of combustion
irreversibility assuming that (i)all the fuel and air mixed in the first step, before reaction,
(ii) fuel oxidation occurs progressively in a number of discrete stages of reaction; each

First Second
I Incremental : Incremental !
I Extent of I Extent of I

r
I
Reaction

Fuel

Air

Chamber 1 Chamber 3 Chamber 5

1.
-I
Global Reactor
I
Combustion
Products Under

I
Chemical
Equilibrium
-
Conditions
Chamber 2 Chamber 4

FIGURE 8 Hypothetical Combustion Chamber for Process Path 2.

 COMBUSTION IRREVERSIBILITY

stage consists of a two-step subprocess: (a) fuel oxidation, and (b) internal thermal
energy exchange. These subprocesses are repeated until the required exit equilibrium
product gas state is attained. Compared with Path 1, Path 2 allows, amongst other
things, the determination of the exergy loss due to mixing of reactants. The boundary
conditions for chamber 1 in which mixing takes place (Fig. 8) are: (1) the incoming
matter rate, composition, and temperature for both the fuel and air, (2) adiabatic
boundaries, (3) no chemical reactions, and (4) the gas constituents exit as a uniform
mixture. The boundary conditions for the fuel oxidation chambers (2,4, etc. Fig. 8) are:
(1) adiabatic boundaries, (2) the gas constituents exit as a uniform mixture, (3) & ;
oxidation product gas constituents (their quantity computed from the increment in
extent of reaction) exit under complete adiabatic reaction conditions, and (4) the gas
constituents not involved in the oxidation exit with the same temperature as that when
entering the chamber.
Finally, following the fuel oxidation, the internal thermal energy exchange subpro-
cess occurs (in chambers 3,5,etc., Fig. 8). The boundary conditions for these chambers
are: (1) adiabatic boundaries, (2) uniform mixture of gas constituents, (3) thermal
equilibrium prevails between all system gas constituents at the exit.
Following the first incremental extent of reaction, the gas components flow into the
downstream chambers, repeating these steps of combustion for the subsequent in-
cremental extent (s) of reaction. This procedure is repeated until the gas constituents
reach the fuel ignition temperature, whereupon the fuel oxidation is assumed to be
instantaneous (i.e., all the remaining fuel is oxidized immediately at the fuel ignition
temperature). The final products of combustion then exit the reactor under the required
global system equilibrium conditions.
As in the Path 1 study, an analysis was first performed to determine the sensitivity of
the results to the chosen number of increments. The results of this analysis were similar
to those shown in Figures 4 and 5 describing results of the Path 1 analysis. Two
hundred increments were chosen for the Path 2 analysis, amply adequate to ensure
results independent of increment number.
Qualitatively similar to the results of the Path 1 analysis, the largest exergy
destruction, 66% to 73% of the total, takes place during the internal thermal energy
exchange (chambers 3,5, etc., Fig. 8). The fuel oxidation (chambers 2,4, etc., Fig. 8) is
responsible for 18% to 25% of the exergy destruction. The mixing process consumes
about 8% to 10% of the total exergy destruction. The corresponding subprocess
exergetic efficiencies, which reflect these results, are displayed in Figure 9.

1 Path 3
In this process path, described in Figure 10, we study the breakdown of combustion
irreversibility assuming that (i) the fuel and air are internally preheated to the ignition
temperature, (ii) the fuel is oxidized instantaneously at the ignition temperature, and
(iii) fuel combustion takes place in the subprocess order of (a) internal reactant
preheating, (b) reactant diffusion/fuel oxidation, (c) internal thermal energy exchange,
and (d) product mixing. Thus, with this scheme, the global combustion process is
envisioned to proceed as follows.
 54 . W. R. DUNBAR AND N. LIOR

1 . 1 . 1 . 1 .

-
J---Oxidation

Mixing

Internal
Thermal
Internal Fuel Energy Product
Preheat Oxidation ~xch@e Mixing
I - - I
.I I 4

Combustion
Products Under
Chemical
Equilibrium
Conditions
Chamber 1 Chamber 2 Chamber 3 Chamber 4

FIGURE 10 Hypothetical Combustion Chamber for Process Path 3.

As the air and fuel enter the reactor, these reactants are internally preheated (by
radiation from the hot combustion products contained in chamber 3, for example).
Upon reaching the ignition temperature, the fuel is oxidized instantaneously. The
instant following this exothermic reaction, the product molecules transfer energy to the
neighboring non-reacting constituents (such as excess oxygen, and N,) in the internal
thermal energy exchange subprocess. Finally, the oxidation products and the non-
' *s--
-
 COMBUSTION IRREVERSIBILITY 55

reacting gas constituents mix uniformly. The boundary conditions for chamber 1 are:
(I)the incoming matter rate, composition, and temperature for both the fuel and air, (2)
no chemical reaction, and (3) both the fuel and air exit the chamber at the ignition
temperature.
In chamber 2, the fuel and oxygen react instantaneously. The boundary conditions
for this chamber are: (1) product gas exits at a temperature characterized by a com-
pleted reaction in an adiabatic chamber, (2) the amount of unreacted fuel and the
oxygen-depleted air exit at the same temperature as when entering chamber 2, and (3)
adiabatic boundaries.
In chamber 3, internal heat transfer (but no mixing) is allowed to take place (i)
between the gas particles in the separated compartments of this chamber, and (ii)
between the gas in this chamber and the lower temperature fuel and air in chamber 1.
Thermal equilibrium prevails between all constituents at the chamber 3 exit.
Finally, in chamber 4, all the constituents mix uniformly. Thus, the two additional
boundary conditions for this chamber are: (1) adiabatic boundaries, and (2) uniform
composition and temperature at the chamber exit, the exiting gas leaving under the
required global chemical equilibrium conditions.
Qualitatively similar to the results of the studies performed using Paths 1 and 2, the
internal thermal energy exchange subprocess is again responsible for the majority of
the combustion irreversibility. According to the results of this study, shown in
Figure 11, this subprocess, which includes the effects of internal preheating, destroys
approximately 74% to 80% of the total exergy destroyed in the process. The fuel
oxidation destroys about 12% to 16% of the total exergy destroyed, and the mixing
process is responsible for approximately 8% to 10% of the combustion exergy
destruction.

-
Energy Exchange

9 20000 20000
wS Oxidation

0 20 40 60 80 100
Excess Air (%)

FIGURE 1 1 Hydrogen Combustion Subprocess Exergy Destruction, vs. Excess Air, Path 3.
 56 W. R. DUNBAR AND N. LIOR

Path 4
In this process path, described in Figure 12, we study the breakdown of combustion
irreversibility assuming that (i) the fuel and air are internally preheated to the ignition
temperature, (ii) the fuel is oxidized instantaneously at the ignition temperature, and
(iii) fuel combustion takes place in the subprocess order of (a) reactant mixing, (b)
internal reactant preheating, (c) fuel oxidation, and (d) internal thermal energy ex-
change. The global combustion process is thus envisioned to proceed as follows.
As the air and fuel enter the reactor, the reactants mix uniformly. Although internal
preheating is occurring at the same time, the internal preheat process is analyzed
separately in order to quantify the irreversibility associated with these two sub-
processes. Thus, after the mixing, the reactants are internally preheated to the ignition
temperature. Following the mixing and internal preheating processes, the fuel is
oxidized instantaneously at the ignition temperature. Finally, the internal thermal
energy exchange subprocess occurs, wherein the products then exit the reactor under
thermal and chemical equilibrium conditions.
As shown, in chamber 1, all the reactant gas constituents mix uniformly. The
boundary conditions for this chamber are: (1) the incoming matter rate, composition,
and temperature for both the fuel and air, (2) no chemical reactions, (3) the gas exits with
a uniform composition at ambient temperature and pressure, and (4) adiabatic
boundaries.
In chamber 2, the fuel and air are internally preheated to the ignition temperature,
The source of heat is the hot product gas in chamber 4, downstream of the fuel
oxidation chamber. The additional boundary conditions for this chamber are: (1) no
chemical reactions, and (2) both the fuel and air exit the chamber at the ignition
temperature.
In chamber 3, the fuel and oxygen react instantaneously. The boundary conditions
for this chamber are: (1) the oxidation product gas molecule exit at a temperature
characterized by a completed reaction in an adiabatic chamber, (2) the unreacted fuel

Internal
Thermal
Reactant Internal Fuel Energy
Mixing Preheat &idation Exchange
t t

I1
1L
Fuel
Feed Qplehcat Qpreheat

* + * *
--Air
Feed
Qpleheat Qprekat
Global Reactor
Combustion
Products Under
I\ !I
Chemical
Equilibrium
- -
/
Conditions
Chamber 1 Chamber 2 Chamber 3 Chamber 4

,*._- FIGURE 12 Hypothetical Combustion Chamber for Process Path 4.

 COMBUSTION IRREVERSIBILITY 57

and air, and the inert gases exit at the same temperature as when entering chamber 3,
and (3) adiabatic boundaries.
Finally, in chamber 4, internal heat transfer occurs (i) between the gas constituents in
this chamber, and (ii) between the gas in this chamber and the lower temperature fuel
and air in chamber 2. The product gas exits chamber 4 under thermal and chemical
equilibrium conditions.
Again qualitatively similar to the results of the studies performed using Paths 1 , 2
and 3, the internal thermal energy exchange subprocess is responsible for the majority
of the combustion irreversibility. According to the results of this study, shown in
Figure 13, this subprocess, which includes the effects of internal preheating, destroys
approximately 74% to 80% of the total exergy destroyed in the process. The fuel
oxidation destroys about 12% to 16% of the total exergy destroyed, and the mixing
process is responsible for approximately 8% to 9% of the combustion exergy destruc-
tion.
To observe the effect of fuel type on combustion subprocess irreversibility, the Path
4 combustion process was reevaluated using methane as fuel, Qualitatively, the results,
shown in Figure 14, are similar to those obtained from the hydrogen combustion
breakdown analysis. The internal thermal energy exchange subprocess is responsible
for the majority of the combustion irreversibility. This subprocess, which includes the
effects of internal preheating, destroys approximately 57% to 67% of the total exergy
destroyed in the process. The fuel oxidation destroys about 30% to 40% of the total
exergy destroyed, and the mixing process is responsible for about 3% of the combus-
tion exergy destruction.
The main difference between the hydrogen and methane combustion exergy destruc-
tion breakdown results is that the methane oxidation subprocess destroys a fraction of
the overall exergy loss which is about 2.5-fold larger than that destroyed in hydrogen
combustion, and the destruction of exergy in the internal energy exchange and mixing
are smaller.

1 . 1 . 1 . 1 .

Internal Thermal

20000
W
8
idation

Mixing
" 0' 0 * 20 40 *Air 60
Excess (Oh) 80 ~ 100 o

FIGURE 13 Hydrogen Combustion Subprocess Exergy Destruction, vs. Excess Air, Path 4.
 58 W. R. DUNBAR AND N. LIOR

g 100000 - Fuel Oxidation

- 100000
w"
J

Mixing
0 I . I . I . , . 0
0 20 40 60 80 100
Excess Air (%)

FIGURE 14 Methane Combustion Subprocess Exergy Destruction, vs. Excess Air,,Path4.

ADDITIONAL DISCUSSION

Comparison of the results from the four process paths analyzed

The results of the four different hypothetical process paths have revealed that the
internal thermal energy exchange subprocess is responsible for more than 213 of the
global exergy destruction. This is shown in Figure 15, as a function of excess air.

FIGURE 15 Exergy Destruction due to Internal Thermal Energy Exchange (for Hydrogen Combustion)
vs. Excess Air.
 COMBUSTION IRREVERSIBILITY 59

The eflect of pre-mixing .

In the analysis of Paths 1 and 3 it was assumed that full mixing occurs after reaction,
while in Paths 2 and 4 pre-mixing was assumed. As shown in Figure 15, no perceptible
differences in exergy destruction due to the mixing subprocess were found between
these modes of mixing.

The efSect of internal preheat

As discussed earlier, internal preheating is the process wherein fuel and air are
preheated within the combustion chamber prior to reaction by heat transfer from hot
upstream products of reaction via radiation and possibly conduction in the combustor
walls.
To explain the effect of internal preheat and using Figure 15, consider first the results
from Path 1 and Path 3, which differ primarily by the fact that Path 3 has internal
preheat while Path 1 does not. In path 1 it is assumed that the fuel is oxidized in
increments until the ignition temperature is attained, whereupon the remaining fuel is
oxidized. With this scheme, approximately 10-25% of the fuel is oxidized prior to
reaching the fuel ignition temperature. In th$ path, 72%-77% of the total combustion
exergy destruction occurs due to the internal thermal energy exchange subprocess.
Path 3, on the other hand, assumes the identical sequence of subprocesses, except
that, prior to any oxidation, the reactants are heated to the ignition temperature, a t
which the fuel is then oxidized instantaneously. The results of this scheme disclose that
the internal thermal energy exchanged and internal preheating are responsible for
approximately 74%-80% of the total combustion irreversibility. Consequently, inter-
nal preheating to the ignition temperature prior to reaction causes the internal heat
transfer irreversibility to increase by about 2%-3% of the total combustion irreversi-
bility. A similar comparison between the results of Paths 2 and 4 (Fig. 15) reveals that
this internal preheating raises the internal heat transfer exergy destruction by approxi-
mately 7%-8% of the total combustion irreversibility.

CONCLUSIONS AND RECOMMENDATIONS

The understanding of the sources of combustion irreversibility and its underlying

reasons was improved by employing a simplified method which does not require the
solution of the spatial Navier-Stokes, energy and reaction kinetics equations. This
method is approximate, since the overall exergy destruction is decomposed into
hypothetical subprocess contributions, and the process is computed along several
prescribed process paths. Nevertheless, useful and consistent results are obtained from
the analysis: independent of the process path selected here, the major contribution to
the destruction of useful energy occurring in typical gaseous hydrogen or hydrocarbon
fuel combustion is likely due to the internal thermal energy exchange (heat transfer)
between the particles within the system.
To reduce entropy production during combustion, this conclusion and the inevita-
bility of this internal thermal energy exchange, point to the need to seek means for
60 W. R. DUNBAR AND N. LIOR

reducing the amount of conversion of the reactants' "chemical energy" to the form of
"thermal energy" which causes this undesirable internal thermal energy exchange.
Recognizing this problem, Keenan (1941) discussed the concept of "reversible
combustion", theoretically attained by preheating the reactants to the equilibrium
temperature and partial pressures without allowing reaction; at these conditions the
tendency toward chemical equilibrium has vanished. Subsequent gradual and revers-
ible alteration of the temperature and pressure of the mixture, following the path of
most stable states, he posited, would lead to increasingly greater combination of the
reactants, resulting altogether in reversible combustion. This concept was expanded
further by Obert (1948, 1973)and Beretta et al. (1992),all, however, acknowledging the
practical difficulties associated with this conceptual procedure, such as the need to
prevent reaction during the initial elevation to the reaction equilibrium conditions, the
high temperatures at that state, and the need to have a reversible chemical combination
process subsequently. Richter and Knoche (1983) proposed a conceptual approach to
attain such reversible combustion by using a metal oxide as an intermediate material
which would combine reversibly with the fuel, and then be restored to its original
condition by a reversible reaction with air. Another way, employing fuel cells, is being
explored by the authors (cf. Dunbar, 1983; Dunbar et al., 1991). Described briefly, fuel
oxidation performed in fuel cells produces useful work (electricity)during the process,
thus generating less internal thermal energy and entropy between the process end-
states, and resulting in a significantly more efficient process. All these are good
examples of how exergy analysis leading to fundamental understanding of process
irreversibilities can point to the development of more efficient practical processes.

ACKNOWLEDGEMENT

The partial support of this study by the Society of Automotive Engineers and the University of Pennsylvania
to the first author (WRD) is gratefully acknowledged. He is currently Vice President of Engineering, Cleaver
Brooks, Milwaukee, WI 53224.

REFERENCES
Arai, N., Hasatani, M., Ninomiya, Y., Churchill, S. W. and Lior, N. (1986).A comprehensive kinetic model of
char NO formation during the combustion of a single particle of coal char. Proc. Combustion Institute
21st Symp. (International) on Combustion, The Combustion Institute, Pittsburgh, PA. pp. 1207-1216.
Arpaci, V. andd Selamet, A. (1988). Entropy production in flames. Combust. Flame, 73,251-259.
Babcock and Wilcox (1978). Steam, its Generation and Use. The Babcock and Wilcox Co., New York.
Bejan, A. (1979). A study of entropy generation in fundamental convective heat transfer. ASME J. Heat
Transfer, 101,718-725.
Beretta, G. P., Lezzi, A. M., Niro, A. and Silvestri, M. (1992).On the concept of a reversible flame. Flowers '92
Florence World Energy Research Symp., pp. 165-177.
Buckmaster, J. D. and Ludford, G. S. S. (1982). Theory of Laminar Flames. Cambridge University Press,
Cambridge.
Dunbar, W. R. (1983). Computer Simulation of a High-Temperature Solid-Electrolyte Fuel Cell, M. S.
Thesis, Marquette University, Milwaukee.
Dunbar, W. R. and Lior, N. (1990). A Breakdown of the Exergy Losses in Combustion, Proc. World Energy
con$, Florence, Italy, Pergamon Press, Oxford. pp. 347-358.
Dunbar, W. R., Lior, N. and Gaggioli, R. A. (1991). Combining fuel cells with fuel-fired power plants for
improved exergy efficiency, Energy, 16,1259.
Dunbar, W. R., Lior, N. and Gaggioli, R. A. (1992).The component equations ofenergy and exergy, ASME J.
Energy Resources Technology, 114,75.
Gaggioli, R. A. (1961). The concept of available energy. Chem. Engng Sci., 16,87.
 COMBUSTION IRREV

Gaggioli, R. A. (1962).The concepts of thermodynamicfriction, thermal available energy, chemical available

energy, and thermal energy. Chem. Engng Sci., 17,523.
Gaggioli, R. A., Yoon, J. J., Patulski, S. A., Latus, A. J. and Obert, E. F. (1975). Pinpointing the real
inefficienciesin power plants and energy systems. In Gaggioli, R. A. (Ed.). Proc. Amer. Power Conf:
Washington, D. C . pp. 671-679.
Gunich, L. V. and Veyts, I. V. (1989). Thermodynamic Properties of Individual Substances. Hemisphere
Publishing Corp., New York.

Obert, E. F. (1948). Thermodynamics, McGraw-Hill, New York, p.459.

Obert, E. F. (1973). Internal Combustion Engines and Air Pollution, Harper and Row, New York, p. 84.
Poulikakos, D. and Johnson, J. M. (1989). Second Law Analysis of Combined Heat and Mass Transfer
Phenomena in External Flow. Energy, 14,67-73.

Second Law Analysis, Ch. 3, ACS Symp. Ser. 235, Washington, D. C.

San. J. Y., Worek, W. M. and Lavan, Z. (1987). Entropy generatiqn in combined heat and mass transfer. Int.
J . Heat Mass Transfer, 30,1359-1368.
Sonntag,R. E. and Van Wylen, G. J. (1982).Introduction to Thermodynamics, Classical and Statistical. Wiley,

Vargaftik,N. B. (1975). Tables on the thermophysical properties of liquids and gases in normal and dissociated
states. 2nd Ed. Hemisphere, Washington, DC..
Wark, K. (1977). Thermodynamics. McGraw-Hill, New York.